Re: [PATCH v7] lib: optimize cpumask_local_spread()

2020-12-10 Thread Shaokun Zhang
Hi Dave, Apologies for the late reply. 在 2020/12/1 1:08, Dave Hansen 写道: { - int cpu, hk_flags; + static DEFINE_SPINLOCK(spread_lock); + static bool used[MAX_NUMNODES]; >>> >>> I thought I mentioned this last time. How large is this array? How >>> large would it be if

Re: [PATCH v7] lib: optimize cpumask_local_spread()

2020-11-30 Thread Dave Hansen
>>> { >>> - int cpu, hk_flags; >>> + static DEFINE_SPINLOCK(spread_lock); >>> + static bool used[MAX_NUMNODES]; >> >> I thought I mentioned this last time. How large is this array? How >> large would it be if it were a nodemask_t? Would this be less code if > > Apologies that I forgot

Re: [PATCH v7] lib: optimize cpumask_local_spread()

2020-11-26 Thread Shaokun Zhang
Hi Dave, Apologies for later reply. 在 2020/11/21 1:48, Dave Hansen 写道: > On 11/17/20 6:54 PM, Shaokun Zhang wrote: >> From: Yuqi Jin >> >> In multi-processor and NUMA system, I/O driver will find cpu cores that >> which shall be bound IRQ. When cpu cores in the local numa have been >> used up,

Re: [PATCH v7] lib: optimize cpumask_local_spread()

2020-11-20 Thread Dave Hansen
On 11/17/20 6:54 PM, Shaokun Zhang wrote: > From: Yuqi Jin > > In multi-processor and NUMA system, I/O driver will find cpu cores that > which shall be bound IRQ. When cpu cores in the local numa have been > used up, it is better to find the node closest to the local numa node > for performance,

[PATCH v7] lib: optimize cpumask_local_spread()

2020-11-17 Thread Shaokun Zhang
From: Yuqi Jin In multi-processor and NUMA system, I/O driver will find cpu cores that which shall be bound IRQ. When cpu cores in the local numa have been used up, it is better to find the node closest to the local numa node for performance, instead of choosing any online cpu immediately. On